change will perhaps account for the fact that the river Oder has two ancient names, Sums and l'iadrus, which have been the cause of much confusion in the geography of ancient. Germany. Indeed the mouth of the river is still called Strineerande.
12. Sp is interchangeable with pa, sk with ks, and ad with ds. For the last we need only refer to the Doric use of eta for (. Instances of the second interchange occur occasionally in Greek and Latin. f?Zor, the misletoe, is written in Latin riseus; coxes-or, ludicrously put down as a primitive in some lexicons, is of course only the superlative of the preposition for *mos. The Latin mimeo has for its participle mistus as well as mistus (= misclus). The tendency to this interchange accounts too for the form seseenti, for sexcenti is never found in the best manuscripts of the best authors. But the Anglo-Saxon and English afford the most numerous instances of this metathesis. Thus the former language has the double forms raps or rasp, a wasp; apse or asps, tremulous (whence the name of the aspen tree) ; !lapse or hasps, a lock (Grimm, ' Deutsche Grammatik,' p. 251); also /rose or frox, a frog; faces or fixes, a fish ; tusc or tax, a tusk ; asce or axe, cinder ; asejan or axjan, to ask (ibid, p. 256). Hence it will be seen that it is a mere accident if in our own tongue axe and traps have been rejected as vulgarisms in favour of ask and wasp. The provinces still prefer the ks and pa. Thus a Kentish countryman talks of a whips, rather than whisp of hay. May we not in this way establish the identity in name of several of our rivers, as Axe, Esc, Esk, and Usk? 13. 8 is often lost. Inattention to this fact is the cause of much confusion iu-the grammars of the Greek language. Thus the neuter nouns in or must once have had a corresponding er in the genitive, 7fVf7, 7f WOOS, &c., afterwards 7f3.05, yeveor. Hence the retention of the s in the vocatives of proper names formed from neuter nouns of this class, as Am-yeses, 471am:reeves, lanipares. (See 'Journal of Educa tion,' vol. iv., 333.) Above all, the neglect of this letter in the original
(as here assumed) forms of certain present tenses leads to apparent anomalies in the derived forms. Thus from watt(o)to we should have without any irregularity neteAetapevos; from ye u(a)aal without difficulty 7f wt.-Tufa; as well as the Latin pusdus, pus-ks-re; from 8 e(a)w, df0-405, in which the sibilant corresponds, as it so often does, to the guttural in liga-re, dice-re, and the English tight from tie. The Latin language in such cases changes the sibilant into an r; but even this language is not at all unwilling to discard an a, particularly at the end of words, as in the double forms magi, and mope, rideris and ride-e, ipsus and ipse, peer for puerus. Nay, even the neuters of adjectives seem to have lost the final s of the nominative in this way. At any rate yetis is used for a neuter nominative as well as pate. The third person of the Latin perfect may possibly owe its occasional long quantity (perrupIt, Hon; subiit, Hon; , Ovid, &e., &c.) to an older orthography ending in id ; for as the other perfects of the indicative as well as those of the subjunctive and infinitive of the active verb, to say nothing of all the passive perfects, are evidently formed by the addition of the tenses of the verb ease, so perreepistis and perruperunt contain in the two last syllables the almost unaltered forms of eels and aunt, and seem to justify the idea that perrupit is a corruption of perrupist, that is, perrupest. As to form, we might compare this corruption with what we know has occurred in the French subjunctive perfect, fear, fusses, fat, that is, fuss'. The French language al:Zombi in examples of the loss of the sibilant. Thus from the Latin asin us, magiskr, floater, ouadragesinta, are derived, first, mete, maistre, nostre, caresme, and then, according to the modern orthography, due, maitre, noire, earl me, to say nothing of the silent a in such words as mu is, roes, isle, eat, Ice.